156 
REPORT-1847. 
surface, the language is !eft unplaced ; where it is evident, the result Is Ihe 
formation of a correct group. By this means, although allied languages are 
separated, distinct ones are not connected. The only important exception to 
this is in the class called Familh Shiho-DunJtali. From this theGallaaml 
Somauli are excluded, whilst the Shilluck, Dizzeia, Dalla, and Agow languages 
arc included. 
Balbi’s uomeuclatare is faulty in one respect; the application of the term 
Famille Troglo(Ii/te to iht? Uisharye dialects. This assumes an affiliation from 
the Troglodyte of antiquity whidi ought to have been proved. 
The absence of any reference to his authorities is a serious inconvenience 
since in collating liis voc.ibularies with tbo.se of other collc<!tors it is difficult 
to determine whether they are additions to those previously known, or m* 
tracts from them. In some cases the orthography is a guide. In one caic, 
however, and probably in more, it misleads. The iJisharye, Noub and Kensy 
Toeabularies are marked German \ which they arc not. They were collected 
by Burckhardt ; whose name, being Gennaii, misled. 
Balbi knew the works of Lyon, Burckhardt, Denham and Clapnerton, pul>- 
lished between 1817 and 1826. 
The third systematic work, which, embracing much beside, embraces aho 
the languages of .Africa, is Friehards. The date of the first edition is 1826, 
of the second 1841. His ethnological maps art? the proper complement to Dr. 
Prichard s writings. These make him the most geographical of the three in- 
vestigators; a matter of no small value and importance. 
In this department, however, of the subject two ditficultie.*! press upon the 
philologist. 'J’hcy occur more or less in all languages ; but most in those of 
Africa. These are,-—Ist, the use of similar names for diflbrentnations; and. 
Slid, the absence of any geographical notice respecting the districts M’herein 
a particular language is spoken. 
As instances of the first compllcatlou we find names as similar as Somauli, 
lumah, Sowuiel and Sofala applied to<lifferent tribes aud different localities. 
This for the eastern coast of Africa. 
We find a Hamara language in Abyssinia, distinguished from the Ambaric; 
so that the particular Abyssjuiau tribe to which the wunpo of Aga- 
tharcides apphes is uncertain. 
We find Shankalas in the neighbourhood of the Dankali, and both near the 
co^try of Dongola. Yet alt three are distinct. 
e find a Shilha tribe of Berbers in Morocco, and Shilluck negroes in Sen* 
naar ; each unconnected with the other. 
Ve find W oloflPs, hulahs and Feloops ; similar in name, unconnected iu 
most other points. And then we find a Fellah population in Egypt. 
these instances might easily he increased. As it is not likely that trilie* 
m the neighbourhood of each other should call themselves by similar namw 
to anything like the extent hen; indicated, the solution lies in the supposition 
that the term by which they are known is a foreign one. Something of the 
kind, as Slated above, occurs in all languages. It occurs, however, to a very 
extraordinaiT extent in Africa. The remedy to the uncertainty which item 
genders is the habit of always stating whether the name by which a tribe ff 
descended IS native or foreign. In Asiatic philology this practice is welUd* 
hered to in Klaproth’s Asia Polyglotta. 
The secoml complication is easily accounted for. A resident at Cairo, 
collects speeiinens of the languages of Africa; and he 
Africans as he finds in the neighbourhood of bis rest- 
of the* 1 however well ll«i»e Africans may supply him witli spedmens 
language, they give him no information in the way of geographp 
